#graph data

I've been getting numerous inquiries as of late at training and meetups in regards to how to viably display time arrangement information with use cases running from hour level accuracy to microsecond level exactness. In evaluating the different methodologies conceivable, I arrived on a tree structure as the model that best fit the issue. The two key inquiries I ended up asking as I experienced the procedure of building the time tree to unite the time arrangement occasions were, "The way granular do I truly need to make this to productively work with and uncover the time-based information being dissected?" and "Do I have to create record-breaking hubs down to the craved exactness level?" The equalization that should be considered is the instatement and viability of all the time arrangement hubs versus the dynamic creation as time arrangement occasions require their presence and the effect the missing hubs could conceivably have when questioning time arrangement occasions by different date and time ranges.

Beginning the Time Tree

I at last concluded that it would be best to make the hour, minute and second level hubs just when expected to join an occasion into a day. So I developed the work done by Mark Needham in his post Neo4j: Cipher – Creating a period tree down to the day. The fundamental displaying change in this stride was to utilize a solitary CONTAINS relationship going from the higher tree toward the lower tree level to rearrange development all over the whole tree through the utilization of profundity based example coordinating. Also I presumed that for sub-second level estimations it would be best to store the full accuracy (i.e millisecond, microsecond, and so forth) on the occasion hub itself, however interface the occasion to the time tree at the second in which it happened on the grounds that any separating or design examination is unrealistic to be important inside of a second (at any rate for the utilization cases I've been hearing).

Interfacing the Time Series Events

In the time arrangement use cases I've been hearing there are a large number of occasions coursing through the framework over brief timeframes so I needed to locate an intriguing information set of important size to use in accepting the viability of the tree based methodology for demonstrating time arrangement information. There has been an open information development in government the last couple years that has been picking up energy so I started hunting data.gov and looking down different urban areas that have open information entryways. I found the city of Seattle has an information entry at data.seattle.gov, which had a valuable, huge time arrangement dataset containing every one of the 911 flame calls got subsequent to 2010. There were around 400k sections in the CSV containing an Address, Latitude, Longitude, Incident Type, Incident Number and Date+Time.

Subsequent to downloading the time arrangement CSV, I expected to a do two or three things to get the information into an all the more well disposed and finish state for stacking into Neo4j.

1. I expelled all spaces from the section names.

2. I needed a millisecond timestamp connected with every line notwithstanding the long UTC content arrangement so I ran a procedure to embed a second time segment that changed over the Datetime string worth to milliseconds and set that esteem in the new DatetimeMS segment.

**Michael Hunger has some extremely valuable tips on utilizing LOAD CSV with Cyhper as a part of his post LOAD CSV into Neo4j rapidly and effectively, which is a beneficial read before you start importing your own CSV information.

Time Series Graphgist

I've incorporated an intuitive time arrangement graphgist beneath with remarked figure for the era and a couple investigation inquiries. You'll see that I'm just stacking the initial 1k lines here in this post and that is just to keep the information set sufficiently little to handle and render quickly for this case. In my own particular testing and exhibit for the meetup bunch I'd stacked and joined all the occasion information. In case you're keen on testing out the bigger dataset or one of your own, I'd be upbeat to offer assistance.

Some accommodating clues for cooperating with the diagram perceptions:

1. Hold move and look up to zoom in.

2. Hold move and look down to zoom out.

3. Hold move and click and drag to move the diagram around the showcase range.

4. Double tap to zoom in around a particular point.

Free talk: Neo4J User Group - Introducing Py2neo and Geoff. Jun 27.

In this presentation for the Neo4j User Group, Nigel Small describes how Py2neo evolved, and provides an introduction to how it is used. Nigel also explores Geoff, a textual format for storing and transmitting graph data (with a syntax heavily influenced by Neo4j's Cypher language).